* rm deprecated and add new sigmoid loss

* fix bw and header file

* cp broadcast div to broadcast mod

* add broadcast mod in op_conf

* mod op

* modify broadcast mod grad

* test broadcast mod

* add math.mod

* replace div to mod for test

* add __mode__ %

* add test mod

* test div instead of mod ok, TODO mod kernel

* floor mod forward test ok

* remove usless lines

* rm broadcast mod grad files

* remove mod grad op in op_conf

* rename mod to floor_mod

* rename mod to floor mod

* format

* of format

* syn test_broadcast_op.py

* loss instance num time shape

* refine

* Update tensorflow cmakefile

* Update setup.py

* Add user op related proto

* Add OpRegistration (#2424)

* Add uncompleted op registration for cooperation review

* Fix the compile bugs

* Refactor the implementation macro of op_reg arg member funcs

* Move impl of Attr() with default to cpp and specialize it

* Add LookUpInOpRegistry()

* Add UserOp as placeholder

* Rename in->input and out->output

* Fix the missing ctor of OpRegistrationBuilder

* Add GetAllRegisteredUserOp() for debug

* Add Log for every user_op registration

* Add const qualifier for Builder::Build() and user_op namespace for REGISTER_USER_OP macro

* Remove the LOG() from ctor of op registrar due to segment fault (maybe a glog bug)

* add customized dir (#2425)

* add customized dir

* customized/.keep

* Add map<string, ListString> output; to UserOpConf (#2426)

* Substitute std::function<...> with alias name and Set default val for those function (#2428)

* Add Kernel Registry (#2431)

* Add Kernel Registration

* Make REGISTER_USER_OP/KERNEL macro available when not in namespace of oneflow

* Add missing TODO of CreateFn parameter

* Add OpKernel for user op

* Fix a little code style

* Add GradRegistry (#2433)

* implement of user_op, instead of get sbp sign (#2429)

* Add UserKernel and UserKernelConf (#2438)

* Add VirtualGenKernelConf() for UserOp and fill UserKernelConf

* Fill KernelRegCtx

* Fix typos and bugs

* Add UserKernel

* Add KernelRegContext(const KernelConf&) as the ctor

* Dev cc python user op conf builder (#2435)

* user op builder in python

* user op wrapper

* Implement of add default vale and check valid between c++ and python

* remove notes

* fixbug and runnable for UserOp complie and add sample as ccrelu

* fix func and class name

* check attr type in op def and op conf; refine code for review

* Dev cc infer tmp size (#2445)

* Refine some code and interface (#2447)

* Add InferContext and Infer Util functions

* Add framework.h as the only included header for user code

* Fix the Dtype infer bug

* Fix duplicated ret

* Fix the KernelRegistration of UserKernel

* Update cc_relu_op.cpp to use InferContext

* Refine and Add test relu kernel

* Add user_op::Blob

* Update InferContext to ptr from const ref

* Add user_op_conf into InferContext and Attr()

* Move cc_relu_op.cpp to customized/ops/

* Add Shape and Dtype into Blob

* Fill the real ReluKernel with Gpu and Float

* Remove unused files

* Add unique_names_ for op registration

* Refactor AttrSeq for re-used of attr function specialization (#2452)

* Refactor AttrSeq for re-used of attr function specialization

* Remove Serialize interface

* Dev cc gen bw user conf (#2449)

* refine python print err

* interface of UserOpWrapper UserOpConfWrapper UserOpConfWrapperBuilder

* implement of user op conf builder in c++

* generate backward op conf for user op

* define grad registration value func

* refine code for review

* check input valid when query need grad

* refine name

* implement of demo ccrelu_grad and test pass of alexnet

* refine ccrelu python

* Add UserOpDefWrapper; Fix .py bug; Add TestReshape op/kernel (#2454)

* Add UserOpDefWrapper

* Update paras of CheckAttrs() from UserOpDef to UserOpDefWrapper

* Fix bug in user_op_builder.py

* Add TestReshape Op&Kernel

* fix ccrelu op grad register and shape infer; Add ccrelu alexnet test python script (#2457)

* Move UserOpConf from op_conf.proto to user_op_conf.proto

* Add test_reshape.py

* Refine the imple of access to attrs in user_op_conf.cpp

* Refactor the way to access AttrVal with AttrValAccessor

* Add GetAttr() in KernelContext

* Rename op_infer_util.h to infer_util.h and Fill paras of InferTmpSize with InferContxt

* Refactor InferContext to simpley user_op

* Refine customized test kernels to get along with interface update

* Dev merge from dev python (#2465)

* Clear session (#2416)

* oneflow.clear_default_session

* fix bugs in oneflow.config.machine

* refactor function return type (#2417)

* fix for py2 (#2418)

* blob parallel conf

* Pr watch scope (#2419)

* pr oneflow.watch*

* merge more code to pass watch_scope.py

* TODO: input_blob_def.parallel_conf

* oneflow.cluster (#2423)

* oneflow.cluster

* no alias for oneflow.cluster.*

* mv cpp_logging_conf from config_proto to cluster_proto

* rename: cluster => env

* rename: Environment => Session

* Free port (#2427)

* oneflow.cluster

* no alias for oneflow.cluster.*

* mv cpp_logging_conf from config_proto to cluster_proto

* rename: cluster => env

* rename: Environment => Session

* auto find a free port for single node environment

* localhost only

* Dev single processor test (#2430)

* oneflow.cluster

* no alias for oneflow.cluster.*

* mv cpp_logging_conf from config_proto to cluster_proto

* rename: cluster => env

* rename: Environment => Session

* auto find a free port for single node environment

* localhost only

* single process test

* Cluster::WorkerLoop

* delete unnecessary OF_BARRIER_ALL

* no longer fork children processes to run tests

* robust contextmanager for CurJobConf (#2434)

* fix of_pure_proto_dir (#2439)

* Ctrl between optimizer (#2443)

* add ctrl edges between optimizors

* update docker file

* sequentialize all optimizors

* Revert "fix of_pure_proto_dir (#2439)" (#2446)

This reverts commit 5031cc86.

* Oneflow unittest (#2448)

* oneflow.unittest.*

* oneflow.unittest.register_testcases

* rename: oneflow.unittest.register_testcases -> oneflow.unittest.register_test_cases

* Test bert inplace with xinqi (#2450)

* update bert script

* update watch_scope test script

* update for debug

* update for debug

* update debug script

* test_inplace.py

* no reshape

* debug IsLbiAllConsumersReachable

* fix inplace

* rm useless code

* update config

* fix critical_section (#2453)

* Patch distribute (#2456)

* backup

* fix bugs

* test_inplace

* Fix InplaceActor when no one consume inplace out regst (#2458)

* disable mutable inplace edge to variable (#2459)

* Fix unittest import conflict for py2 (#2460)

* Create __init__.py (#2464)

* update ccrelu_alexnet.py

* Update code that use UserOpConf to UserOpConfWrapper

* Add paras of GetSbp

* Refactor KernelContext with pure virtual member function

* Refactor KernelRegContext with pure virtual member function

* Refactor UserKernelContext ctor to simplify code

* Refactor InferContext with pure virtual member function

* Rename Blob to Tensor; BlobDef to TensorDesc

* Remove unused log and file

* Dev cc user op sbp (#2480)

* ccrelu multi-gpu runnable

* example of ccrelu op get sbp sign

* sbp context

* add example for get sbp using LogicalTensor...

* GetSbpFnUtil::MirrorSplitAtDim0

* fix bug of useless

* Fix the bug of init UserKernelContext (#2487)

* Fix the bug of init UserKernelContext

* Fix due to comment

* Refine code

* Add missing SetCheckAttrFn implementation

* refine python user_op_builder.SetAttr(); refine reshape test an… (#2505)

* Make UserOp runnable; refine user op python test for new change in dev_python
Co-authored-by: Ncheng cheng <472491134@qq.com>